(194) Performance of the Situational Awareness Scoring System (SASS) in Adult Cardiovascular ICU

Introduction: The Situational Awareness Scoring System (SASS) is a new tool that automatically collects clinical data, including vital signs, vasoactive medications, respiratory support, and blood gas analyses from electronic health records. It processes this data through an algorithm to calculate a real-time score reflecting an intensive care unit (ICU) patient’s clinical state. This study assessed the SASS's performance in the adult cardiovascular ICU (CVICU).

Methods: This retrospective cohort study examined patients admitted to the Cleveland Clinic adult CVICU from March 2015 to December 2021. Patients who were already on extracorporeal membrane oxygenation (ECMO) at the time of CVICU admission or had Do Not Resuscitate orders were excluded. The primary event was a combined endpoint of cardiac arrest (CA) or the need for ECMO. SASS score summary statistics including mean, median, and maximum were calculated within 1, 2, 4, 6, and 8-hour intervals, counting backward from the time of the event for event patients or the end of the CVICU stay for control patients. Analysis was performed at the interval level. For each interval length, an analysis dataset was selected which included all intervals ending in an event and a 1:1 random selection of control intervals. Cross-validated discrete time logistic regression models, using SASS score summary statistics to predict events, were employed to calculate the areas under the receiver operating characteristic (AUC) curves and odds ratios (ORs).

Results: A total of 37,490 CVICU admissions in which 302 events occurred were analyzed. Maximum SASS score within 6-hour time interval achieved the highest AUC of 0.91 [95% Confidence Interval (CI): 0.88-0.93] of all assessed summary statistics and time intervals (Figure). A cutoff value of 45 for maximum SASS score within 6-hour time interval resulted in sensitivity and specificity of 84% and 84%, respectively, with an OR for event of 1.5 [95%CI: 1.4-1.6] for every 10 unit increase in the maximum SASS score.

Conclusions: The SASS demonstrated satisfactory discriminatory performance between patients with and without CA or the need for ECMO. It shows promise as a situational awareness tool in the CVICU. Future prospective validation studies are recommended.